treme sinuosities, and which at the same time gives a good "fit" as judged by some arbitrary criterion-say, the mean square error. These requirements, however, are mutually destructive, and the statistician must exercise his judgment, in view of all the known facts, in effecting a happy compromise between these requirements. The whole question has been well stated by E. C. Rhodes: Given a series of points which require to be smoothed, we have always before us two principles, (1) that we must obtain a curve without the sinuosities which are not inherent in the data, and (2) we must get a curve which "fits" by least squares. These two principles are mutually inconsistent. On the one hand we can obtain the curve which actually goes through the points-here the sum of the squares of the errors is zero, but the curve is composed of perturbations or sinuosities. On the other hand we can obtain a perfectly smooth curve, a straight line, a parabola, a sine curve, or whatever may be chosen to suit the given series, and we obtain a large mean square error. Somehow by careful balancing of these two mutually destructive principles we can obtain for any data a curve which is reasonably smooth and which at the same time gives a series of errors which have a reasonably small mean square error. The smoothness we obtain is not the best, nor is the sum of the squares of the errors a minimum.11 In the case of most statistical series, it is impossible to get an objective test of the wisdom of one's compromise between smoothness and goodness of fit. With the data before us, however, we may get at least a rough check on the suitability of the "compromise curve" for the problem in hand.12 The main, if not the only, object in fitting trends to consumption and prices is to enable us to remove the disturbing factors which affect these variables and mask or hide their true relationship; for we know that when all of the disturbing factors have been eliminated, the correlation between the variables is perfect. Hence those trends are most suitable which best accomplish this object. That is to say, of all the pairs of trends that may be obtained by selecting one trend for consumption and one for prices, that pair 11 Tracts for Computers, No. VI, "Smoothing," pp. 43-44 (Cambridge University Press, 1921). 12 See Journal of Political Economy, XXXIII (October, 1925), 503-4. Hand-to-Mouth Buying. By LEVERETT S. LYON. Washington: The Brookings Institution, 1929. Pp. xv+487. $4.00. In this study, a wealth of statistical and verbal data, both published and unpublished, has been effectively organized and ably interpreted. After a brief but lucid introduction to the problem of hand-to-mouth buying, the order basis (ratio of average unfilled orders to shipments for the ensuing twelve months) is described for a number of single industries and for a number of industrial branches. The size of orders and of shipments is also studied for the business units which made their data available-mostly firms in the canning industry. This being the direct evidence bearing upon the problem, the effects and concomitants are then discussed. Of these, the stock burden (ratio of stocks to shipments or sales) is measured for a number of industries and of branches of trade, and the stability of business is studied with reference both to the flow of goods and to the level of stocks. The data go back to 1919 in a few cases, but most of the series begin after 1921-22. The study is thus confined to the last decade, and it is not surprising that the conclusions are neither general nor definite. Since 1920 there has been a noticeable tendency toward restricted forward ordering, a larger percentage of small orders and purchases, and greater stability in the production, shipments, and stocks. But significant variations appear from one industry to another. It is an admirable feature of the book that it makes much of these differences, and presents the findings in copious detail. If, in consequence, the book is less interesting to the general reader, it is the more valuable to a student interested in the variety of business experience and critical of broad but misleading generalizations. In view of the short period studied, the last part of the book, the discussion of buying policies in the past as revealed by trade-journal records for the drygoods, shoe, and other trades, is very enlightening. It is shown that hand-to-mouth buying existed in the past, and that purchasing policies followed the long swings in the general price level. The general price decline from the 1880's was accompanied by frequent recurrences of hand-to-mouth buying, while this practice was rather exceptional during the years 1900-1920. Other factors affecting the purchasing policies of business men, such as elements of style and means of communication and transportation, are discussed briefly in the last chapter. One regrets that this treatment of factors affecting business buying, especially the treatment of trends in price levels, has not been developed in greater detail. Indeed, the only important criticism to be made is that, while the study presents the evidence in full, it does not sufficiently analyze the determining and conditioning factors in the problem. Thus, the congeries of elements that go to make up the cyclical swing all bear upon buying policies, not only upon short-time changes in practice, but also upon certain aspects of trends in practice. The recent trend toward hand-to-mouth buying may be conceived in large part as a change from the violent cycle of 1919-21 to the milder swing of 192124, and to the still milder cycle of 1924-27. This feature does not come out in the discussion, and an impression is created that the ratios and verbal data are presented without full realization of the peculiar and limited character of the period sudied. On the other hand, it is not certain that a deeper analysis could be combined effectively with the full description of the situation, centered, as it is by the author, around the concept of hand-to-mouth buying today. Despite this limitation, the wide range of the material and the admirable organization of the data make the study both valuable to business executives and stimulating to empirically minded economic students. NATIONAL BUREAU OF ECONOMIC RESEARCH SIMON KUZNETS The Small Loan Situation in New Jersey in 1929. By WILLFORD ISBELL KING. Trenton: New Jersey Industrial Lenders' Association, 1929. Pp. 101. $0.25. In 1929, the New Jersey Legislature, by an amendment to the law governing small loan companies, reduced the legal rate on small loans from 3 per cent to 12 per cent a month. The 3 per cent rate had been in effect since 1914, in which year New Jersey enacted a law, almost the first of its kind, regulating the business of making loans of less than $300. Under the 1914 law, the small loan business had developed steadily, and in 1928 there were 429 companies licensed to carry on the business of making small loans, with an investment of about $19,000,000. The reduction in rate which became operative in February, 1930 meant, as has since become evident, that nearly all of the companies would have to cease making new loans and retire from the state. Threatened with extinction, the companies determined to carry their case to the public. Unfortunately for the companies, few influential people knew any thing about the business as no general description of it had as yet been published. The New Jersey Industrial Lenders' Association, the trade organization of those engaged in the small loan business, therefore, engaged Dr. King to make an impartial economic and statistical survey of the business in New Jersey. The expenses of the survey were born by the Association but the investigator was left entirely untrammelled in ascertaining and in reporting the facts. The selection of Dr. King was due to the desire of the Association to obtain a man of such standing that his conclusions would bear weight. As it happens, the writer of this review spent nearly three years investigating this business in the United States for the Russell Sage Foundation and conducted a somewhat similar statistical study which has just appeared in print. There is no question in the mind of the reviewer that Dr. King's study has been made without fear or favor and that the facts presented in his report constitute an excellent picture of the small loan situation in New Jersey. Dr. King's investigations covered twenty-nine companies selected with due regard to geographical location, size, and ownership. In all, 24,846 borrowers were studied. The most surprising thing to many would be the number of people who borrow from these small loan companies. Dr. King estimates that at any one time one married man in eight was a patron of some small loan company. This ratio would of course be increased if one considered the number of those who borrow during the period of a year. The average loan was found to be $169.81. White borrowers on the average pay off their loan in 279 days, colored borrowers in 294 days. Dr. King states "that, in general, the borrower must put in four or five days to pay for the accommodation of having a loan amounting approximately to a month's income, and extending over a maximum period of about nine months." Approximately one-fifth of the borrowers were employers or selfemployed, four-fifths employees. Manual workers including servants made up five-ninths of the borrowers and borrowed about one-half of the money. Colored borrowers were found in excess of their numbers in the population, 12 per cent as against 4 per cent, a situation due no doubt, as the author says, to the fact that the negroes as a class belong largely to the group most likely to borrow. Dr. King takes pains to point out that the borrowers do not constitute, as some seem to believe, the flotsam and jetsam of humanity but are persons with income and ability to repay if the terms are adjusted to their conditions. The most important reason for borrowing is to meet current household and miscellaneous expenses. Nearly one-fifth gave this as the reason for borrowing. The second most important reason given was to consolidate and liquidate existing debts. No information was obtained as to how these debts arose. Dr. King believes that fully one-half of the loans are made for the purpose of paying off other creditors. Occupation naturally has some effect on the reason for borrowing. Onethird of all loans by non-professional business men are made for business purposes. Employees, of course, make few loans for this purpose. The survey included a study of exchanges, set up by the companies, to protect themselves from particularly bad risks and to prevent customers from overloading. Personal interviews with borrowers in their homes were also made to obtain their point of view with respect to the business. With few exceptions, the attitude of the borrowers was distinctly favorable to the companies. They testified to fair treatment and to benefits received from the loan. Incidentally, it may be of interest to remark that the amendment reducing the rate was not repealed in 1930 and that the companies are gradually collecting their funds and closing down. Dr. King's survey clearly shows the importance of credit institutions to which people without bankable collateral or powerful friends can turn in times of emergency. Had the people of New Jersey been fully cognizant of the facts contained in this study it is doubtful if any such ill-considered legislation would have passed. SWARTHMORE, PA. LOUIS N. ROBINSON The Evolution of War. By MAURICE R. DAVIE. New Haven: Yale University Press, 1929. Pp. x+391. $4.00. The author of this book has made a valuable contribution to the study of war through assembling a considerable array of facts and even more quotations under a classification which will enable future investigators to find them. One has before him instances of war-making to increase the supply of human meat, of human sacrifices, and of human slaves; war-making for blood revenge, for trophy heads, and for glory; war-making for land, for cattle, and for women. One can also turn to cases where tribes fight to exterminate each other, where they fight for sport without great damage, and where they fight as a formal rite. |